Vision, Principles and Impact of Reconfigurable Manufacturing Systems

Yoram Koren and Galip Ulsoy

University of Michigan

Powertrain International, Volume 5, Number 3, 2002, pp. 14-21

Obtained from Professor Yoram Koren’s Site: http://sitemaker.umich.edu/ykoren

Direct Link to Paper: http://www-

personal.umich.edu/~ykoren/uploads/Vision,_principles_and_impact_of_reconfigurable_manufacturing_systems.pdf

Vision, Principles and Impact of Reconfigurable Manufacturing Systems

Yoram Koren, Galip Ulsoy • University of Michigan

Figure 1: Basic Foundations of Manufacturing

14 POWERTRAIN INTERNATIONAL

e Introduction

Cost, Quality, and Responsiveness are the three foundations on which every manufactw-ing company stands [Figure 1}. Low Cost has been an eco­nomic goal since the introduction of mass production in the 1920s. High Quality is an economic goal associat­ed with the lean manufacturing para­digm [6], and more recently has been the focus of Six Sigma [5]. Cost-effec­tive Responsiveness is achieved by reconfigurable manufacturing [3].

For manufacturing companies to be competitive in the global economy, these three goats are equally impor­tant. Like a 3-legged stool that needs equally strong legs to be stable, a manufacturing enterprise should strive cowards low cost, high quality and rapid responsiveness.

One basic difference among these par­adigms is that Lean Manufacturing and Six Sigma deal with improving the system operations. By contrast, Mass Production and Reconfigurable Manufacturing focus on new ways of designing production systems as well as operating them.

Reconfigurable Manufacturing

At the heart of a Reconfigurable Manufacturing System (RMS) is a set of core characteristics, as defined below:

Modularity - the comparnnentalization

of operational functions and require­ments into quantifiable units that can be transacted between alternate pro­duction schemes to achieve the most optimal arrangement to fit a given set of needs.

Scalability- the ability to easily change existing production capacity by rear­ranging an existing production system and/or changing the production capacity of reconfigurable compo­nents (e.g., machines) within that sys­tem.

Integrability - the abil ity to integrate modules rapidly and precisely by a set of mechanical, informational, and control interfaces that enable integra­tion and communication.

Convertibility-the ability to easily transform the functional ity of existing systems, machines, and controls to suit new production requirements.

Customization - the ability to adapt the customized (non-general) flexi bility of production systems and machines to meet new requirements with a family of similar products.

Diagnosibility-the ability to automati­cally read the current state of a system and controls so as to detect and diag­nose the root-cause of defects, and subsequently correct operational defects quickly.

When these characteristics are embed­ded in the system design, a high

- -

P R 0 C E S S E S

Reconfigurable Machining System -~~-

Figure 2: Reconfigurable Machining System

degree of reconfigurability is achieved. Furthermore, they enable the RMS to serve as a cost-effective compromise between the low produc­tivity but high flexibility of flexible systems, and the ultra-high productiv­ity but low flexibility of dedicated lines.

The Cost-Effectiveness of RMS

is achieved through:

• Adjustable resomces that enable system scalability in response to changing market demands and sys­tem convertibility to new products of the same part family. Resources may be adjusted at the system level (e.g., adding machines) and at the machine level (changing machine hardware and control software).

• Customized flexibility for a part family that can allow for (1) opti­mal mix between CNC and dedicat­ed machines in a system, and (2) multi-tool operation on a CNC­rype machine, thereby multiplying the productivity of the machine.

The definition of a reconfigurable manufacturing system is, therefore, as follows (3]:

A Reconfigurable Manufacturing System (RMS) is one designed at rhe

ourset for rapid change in its struc­ture, as well as its hardware and soft­ware components, in order to quickly adjust its production capaciry and functionality within a part family in response ro sudden market changes or intrinsic system changes.

Vision and Scope

The vision of transforming the manu­facturing industry in the 21st Century by introducing next generation tech­nology -a living, evolving factory­the reconfigurable manufacrunng sys­tem (RMS}, was the basis of our pro­posal to NSF to form the Engineering Research Center for Reconfigurable Manufacturing Systems (ERC/RMS} that will build the scientific base for responsive, reconfigurable systems (see http://erc.engin.ttmich.edu). The Center, we proposed, would develop RMS design, economic evaluation, and reconfiguration methodologies that could be applied to several man­ufactu-ring domains (i.e., RMS), bnt would implement them just on machining systems (i.e., RmS) in our testbed. This vision has been thor­oughly reviewed during the last sum­mer, and still found to be very com­pelling now.

Responsive systems nre systems whose

production capacity JS adjustable to fluctuations in product demand and whose functionality is adaptable to product variety. The vision of the reconfigurable manufacturing para­digm may be summarized by

Exactly the Functionality and Capadty Needed_ ..

... Exactly When Needed

The ERC/R1\1S inrroduced the con­cept of reconfigurabiliry and is devel­oping a generic reconfiguration sci­ence base to realize rhis vision. The Center's testbed and case studies are primarily focused on reconfigurable machining systems (RMS), which are a special case of RMS that is the Center's overall vision. During the last two years we have broadened the Center research to also include pro­jects related to assembly and semicon­ductor fabrications.

Example of a Reconfigurable Machining System

We would like to give an example of a reconfigurable machining system. Imagine a part a that can be complet­ed in three operations: Op. 10 = milling straight surfaces with a 3-axis horizontal milling machine (CNC 1 0); Op. 20 = drilling of many holes

POWERTRAIN INTERNATIONAL 1s

Figure 3: Arch-Type Machining

Reconfiguration Cost

Reconfiguration Time

Implementing t he RMS Principles ensures that the RMS is reconfigured rapidly and cost effectively

Figure 4: Reconfigurable Time and Cost Relation

Product A

I

Product A+ B

RMS Time 3

Product B+ C

Functionality

Figure 5: Capacity/Functionality Benefits of RMS over OML and RMS

1 c PO WERT RAIN INTERNATIONAL

Multiple Products

perpendicular to the machined sur­faces; Op. 30 = Milling of features, which can be done only after drilling the holes (CNC 30).

One possible system configuration is shown in Figure 2 that shows a 3-stage machining system plus an addi­tional stage in which in-process inspection is performed. CNC A-10 is an identical machine to CNC B-10 and CNC C-10, and all three of these machines can perform Op. 10. Sinlilarly, CNC A-30 is identical to CNC B-30 and CNC C-30. Identical machines (that are loaded and unloaded with a gantry) are added in parallel to adjust the system capacity to the required designed volume of annual parts.

The dr illing of the holes (e.g., 15 holes, some with different diameters) could be done in CNCs 10, but it would be time consunling (position­ing and drilling of 15 holes plus tool changing) which, in turn would require adding more machines in par­allel in Op. 10 to supply the needed demand, and consequently the cost would be increased. Therefore, the drilling operation is separated, and is being performed by a Reconfigurable Machine Tool (RMT) that can drill 15 holes simultaneously. This is a very rapid operation, and therefore only one RMT machine is required in Op. 20, while in the other stages three CNC machines are needed.

If another part ~ has to be also pro­duced on the same system in a batch­type operation, then the system is stopped for a short period (e.g., 20 minutes) tO perform spindle-head change, such that the new hole-pat­tern fits the part geometry. If a ran­dom part production is needed, then an additional RMT-~ has to be added in the second stage, and part routing would be done accord ing ro rhe part type. (The random part production requirement is more expensive.)

. . . 0 I

PROCES S E S o~

RMS Sdence Base System-Level

Process Planning Reconfigurable Machine

Design Theory

Life-Cycle Economic Modeling

System Configuration Rules

Figure 6: RMS Science Base

The part may move in this system through several routes (9 in this example). This has an advantage in cases of machine failure - an alternate route keeps the system running, while in a serial line configuration (with no buffers), for example, each machine failure stops the line. Nevertheless, if the RMT fails, the system stops. One option to keep the system running in such cases is to add a Stand-By gener­al-purpose CNC in the second stage, which provides a safety capacity for this system. T his CNC can be utilized as a stand-by machine not only to Op. 20, but also to Op. 10 and Op. 30, with the appropriate change in the part routing. (The utilization of the stand-by machine would require using different process plans.) One may conclude that more alternate roures for a part to be produced on a system enhances its capability to cope with unexpected events, or in our term, enhances the system reconfig­urabiliry.

In this example we introduced one

ControUer Configuration Methodology

Open-Architecture Principles

Stream-of-Variations Theory

class of RMTs - a class by which the system capacity is adjusted in a cost­effective manner. Another class of RMTs is one in which the functional­ity of the machine may be rapidly adjusted. An example of this class is our Arch-type machine, that al though having only 3 controlled axes, can drill and mill surfaces at inclined angles The angle can be changed dur­ing a short reconfiguration period (3 minutes). The operation of this machine can be demonstrated in our testbed.

RMS Principles

ReconfigurabiJity can be achieved in any production system, if time and cost are not considered. Bur in order to perform cost-effective and rapid reconfiguration, the design and oper­ation of reconfigurable manufactur­ing systems must be based on several RMS principles:

1. To enhance the responsiveness of a manufacturing system, the RMS core characteristics should be

embedded in the entire system as well as in its components (mechani­cal, communications and controls).

2. The RMS contains adjustable pro­duction resources to respond to imminent needs.

• The RMS capacity is rapidly scalable in small, optimal increments.

• The RMS functionality is rapidly adaptable to the pro­duction of new parts.

3. The RMS is designed around a part family, with just the customized flexibility needed for producing all parts of this part family.

4. The RMS contains an economic mix of flexible and dedicated equip­ment, as well as Reconfigurable Machine Tools whose functionality and productivity can be readily changed when needed.

5. Continual monitoring and diagnos­tics is embedded in the RMS mod­ules to enhance its response to a fault or quality/productivity degradation.

POWERTRAIN INTERNATIONAL t7

6. In general, shorter manufacturing systems (i.e., smaller number of stage~) are more reconfigurable, but they require higher investment cost in machine functionality.

7.ln general, systems with a large number of alternative routes to pro­duce a part are more reconfig­urable, but rhey require higher investment cost in the material-han­dling system.

8. The RMS possesses cost-effective safety capacity and stand-by func­tionality th<lt is utilized to cope with unpredictable event').

9.Decision-making capability is embedded in the RMS to reduce its response time to unpredictable events.

10. The organization of the manpow­er that operates the RMS is struc­tured according to the R MS core characteristics and includes people and reams (modules) that are dedi­cared to particular tasks as well as people and reams that are flexible in their assignments.

Figure 7: Principles of Reconfigurability

t a POWERTRAIN INTERNATIONAL

A manufacturing company that pos­sesses responsive production systems designed according to these principles can respond quicldy to short windows of opportunity for new products, and can cope witb unpredictable, high-fre­quency market changes as well as with unpredictable equipment failure in the system.

Reconfiguration Science Base

The ERC/RMS is the world intellectu­al leader in cbe development of funda­mental knowledge concerning recon­figurable systems, and cooperates with other world-wide centers to pro­mote its use in appLication domains other than machining (e.g., assembly and semiconductor fa brication) (Figure 7).

The Center is taking the Lead in devel­oping a new reconfiguration science base, defined as a set of methodolo­gies that enable the implementation of the RMS principles which, in turn, facilitate rapid and economic recon­figuration, The science base will have

a potential widespread impact even beyond manufacturing Consequently, the benefits will not be confined to machining systems, but will be spread throughout the nation's manufactur­ing sector and beyond.

Both the dedicated manufacturing line (DML) and the FMS are static sys­tems. The DML is designed to pro­duce one product at large capacity. The FMS can produce several prod­ucts but, because of its high cost, it is designed for low capacity. The RMS is a dynamic, evolving system, so that its capacity and functionality change in time, as needed (Figure 5).

Goals

We believe that reconfiguration sci­ence will form the basis for a vital pro­duction technology in this era of glob­al market competitiveness - that it will evolve into an entirely new manu­facturing field with enduring benefits for the US economy and society. The RMS Center is the world leader in developing the knowledge and science-

Refine Science

Semiconductor Fabrication

Reconfigurable Manufacturing Systems

- --------------- --

! PROCESSES I

base for this field, including:

• Development of computer-aided design (CAD) teclmiques and mathe­matical roots for the systematic design of manufacturing systems and their economic evaluation,

• Creation of a new generation of reconfigurable machines and recon­figurable controllers, as well as methodologies and software tools for their design.

• Reduction in ramp-up time after installation or reconfiguration of manufacturing systems by develop­ing system root-cause analysis tools and on-line part measurement methodology.

These three components are being inte­grated into a testbed facility for testing and obtaining feedback for refining the science base. In addition, the Center is developing educational materials and courses, so that the knowJedge base for reconfigurable systems engineering can be disseminated to industry and to other universities. Let us elaborate on these points.

CAD for Manufacturing Systems

The most effective way to accelerate the time-to-market for new products is by reconfiguring existing manufacturing systems, wruch must be done rapidly, efficiently, and in a systematic way. The only way to take economic advan­tage of the dramatic reduction of prod­uct development time achieved by com­puter-aided design (CAD) is to develop CAD for the design of the manufactur­ing systems that produce the products as well as for their efficient reconfigura­rion. This task, although more difficult than the development of CAD for products, will have an unprecedented impact on all manufacturing industries, and, in turn, on the economy. Developing CAD for manufacturing systems will be perhaps the most sub­stantial contribution of the ERORMS to science and the national economy.

Economic Evaluation

If we take into account the entire life­cycle cost of a production system, reconfigurable systems may be less expensive than dedicated lines or flexible manufacturing systems. The main factor that makes the RMS less expensive is that, unlike the other types of systems, the RMS is installed with exactly the production capacity and functionality needed, and may be upgraded (in terms of both capac­ity and functionality) in the future, exactly when needed. However, mathematical tools are needed to compute the optimal reconfiguration intervals and to assess the optimal mix of reconfigurable capacity in the manufacturing process.

Reconfigurable Machines

The main components of RmS are CNC machines and Reconfigurable Machine Tools (RMTs) - a new type of modular machine that bas a changeable structure that enables the adjustment of its resources (e.g., adding a second spindle unit) accord­ing to the need. The reconfigurable machine tool (RMT) is a hybrid between a dedicated machine that is designed to machine just one part type with high efficiency, and a CNC machine that has "general flexibility" and can produce a variety of types of parts. The RMT is designed around a part family, and therefore it can utilize programmable, multiple-tool spindles that drill simultaneously, and/or a special geometry to machine a partic­ular part, with adjustment period between different parts of the same part family, thereby providing "cus­tomized flexibility." Adjusting the maclllne tO new parr geometry within the part family requires a short con­version time {e.g., 20 minutes).

Reconfigurable Controllers

We predicted in our original proposal that industrial control wiU evolve in a

manner similar to that of desktop computer hardware and software. There will be companies producing hardware platforms, companies developing industrial control soft­ware, and companies developing generic software tools that can enable users to: (1) integrate their propri­etary control modules into config­urable user-tailored controUers and (2) write and debug their control code for large systems (i. e., replacement of ladder diagrams). Out of this will grow a new software industry of sys­tem integrators. With the aid of these software tools, the control design of new and reconfigured plants will be done systematicaUy, which, in turn, wiU create major savings generated by shorter design time and shorter ramp­up rime for control systems.

Ramp-Up Time

Rapid ramp-up is very important for any manufacturing system since, dur­ing rhis period, following installation, full-volwne production at required quality levels is not achieved. However, ramp-up becomes critical in RMSs, which, by their nature, must go through many processes of recon­figuration and subsequent calibration and ramp-up periods during their life­time. If rapid ramp-up is not achieved, the RMS cannot meet expectations and it is not advanta­geous. The ramp-up period can be dramatically shortened by using two complementary methods: (1) diagnos­tics based on process knowledge and statistics embedded in components and propagating to the system level, and (2) new real-time part dimension measurement technology combined with systematic calibration of machines. Integration of these two methods enables a systematic identifi­cation of the sources of part quality problems through "intelligent" root­cause analysis.

POWERTRAIN INTERNATIONAL 19

Accomplishments

RMS technology requires a systems approach ro creating new knowledge for the design and operation of recon­figurable manufacturing systems. Several elements of this new body of knowledge, whtch we ca ll "the recon­figuration science base," have already been developed.

• Given a part family, volume, and mix, a System-Level Process Planner suggests alternative system configurations and compares their productivity, parr quality, convert­ibility, and scalability options.

• A Life-Cycle Economic Modeling methodology, based on blending dynamic programming with option theory, recommends the system that will be optimally profitable during its lifetime.

• A Reconfigurable Machine Tool {RMT) design methodology allows machines to be systematically designed, starring from the features to be machined on a family of parts. A new arch-type RMT, ·which has been designed and built, forms the basis for a new direction i.n machine research.

• A logic control design methodology for sequencing and coordination con­trol of large manufacturing systems results in reconfigurable ru1d formal­ly verifiable controllers that can be implemented on industrial PLCs.

• A Stream-of-Variations (So V) methodology based on blending state-space control theory with in­process statistics forms a new theo­retical approach for systematic ramp-up time reduction.

• A measurement Scheme Selection methodology for modular machines and large systems ensures that the quality of the finished parts is eco­nomically optimal. (The develop­menr of a new In-process Reconfigurable Inspection Machine is underway.}

20 POWERTRAIN INTERNATIONAL

Broad Impact of RMS: Changing the Relationships between Builders and Users

With RMS deployment, the original supplier (i.e., builder or integrator) of the system will implement also the reconfiguration process and integrate new technology into existing systems. This, in turn, will change the nature of the relationships between equip­ment suppliers and users of RMS technology. It will prompt the devel­opment of enduring strategic relation­ships (rather than one-rime sales) in which the suppliers frequently recon­figure systems for users and continu­ously integrate profitable upgrades, allowing users to rapidly capture ris­ing market demand, comply with gov­ernment regulations and supply better products to customers.

Furthermore, proximity between the plant where an RMS is installed and the location of the machine builder will become extremely important. To enable quick reconfiguration, North American users will have to buy all their machining systems from US builders. The users will have a busi­ness interest in maintaining a strong domestic machine tool industry. Therefore, the new reconfigurable technology may enable the US to once again become a world leader in the machine tool industry.

Commercial Reconfigurable Machines

Several of the ERC/RMS industry members have started to bui ld and advertise their products as Reconiignrable Machines (e.g., Masco Machines in Cleveland, and CELCON - Cellular Concepts in Detroit). Livernois Engineering is seeking our expertise ro jointly devel­op a conceptual design of reconfig­urable assembly machine for car heat exchangers, which their US (e.g., Delphi) and European customers are very excited to see realized. The

February 2002 issue of Mechanical Engineering magazine [1] featured an article describing reconfigurable man­ufacnJring of connecting rods by a small supplier firm, Tri-Way. Our new Arch-Type Reconfigurable Machine will be displayed in the IMTS in Chicago in September, and boost interest in RMTs even more.

US Patents. The Center is protecting several of its inventions by patents. Five US patents have been granted and five new applications have been filed. Three of our patents, "Reconfigurable Machine Tools," "Reconfigurable Manufacturing Systems," and "Reconfigurable Logic Controller" form the cornerstones of designing RMSs. Although the full impact of these patents will only be observed about 10 years from now, we believe that our patent "Reconfigurable Machine Tools," issued in 1998, inspired the creation of a new modular machining center for engine block production by Ingersoll Milling Machines [2].

Center's Software Transfer to Industry. The ERC/RMS has started to deliver beta-version system design software packages to the end-users. ln particular these are software pack­ages aimed at enhancing system pro­ductivity, such as our line balancing algoritlm1 and PAMS (Performance Analysis of Manufacturing Systems). Software for ramp-up time reduction, based upon our stream-of-variation theory, has been developed and vali­dated in the testbed. We are now planning the first in-plant evaluations of this software with Ford and one of its suppliers, Citation.

Stream-of-Variations Impact

The development and implementation of the So V in assembly (supported by NIST/ATP) has made a significant impact in industry. The So V method­ology, however, has not been inlple­mented yet in machining plants.

• --·- ·

Nevertheless, during the last few years approximately 20 papers have been published (or submitted) on various aspects of the So V topic. In 2001 , four of these papers received the BEST Paper award from ASME Mfg Division, ASME Design Division, INFORMS QSR Section, and NAMRC. Furthermore, four former ERC students/researchers (Judy Jin, Yu Ding, Darek Ceglarek, and Shiyu Zhou) who worked on the So V topic have joined or will join the faculty in top induso·ial engineering deparonenrs in the US, which will enhance the deployment of the So V methodology.

Education and Outreach

The essence of reconfigurable systems is design for the future, where the future is unknown and can be only projected. This principle, when applied to engineering education, may open a new approach in solving engineering design problems. We have developed a new undergraduate Manufacturing Systems Concentra­tion at the University of Michigan, with its supporting courses, that aims ar introducing this approach. The ERC helps to attract a talented and diverse group of students to the area of manufacturing. Our outreach to K-12 students specifically targets underrepresented groups, involving Detroit area junior-high-school and high-school students. We also reach out to general public through our museum exhibits.

New Graduate Courses

Our two new graduate courses have begun to achieve national attention. The course "Reconfiguration Science­Base" was taught in Winter 2002 ro 60 students, over half of them at remote industry sites in the US and Mexico. The course "Agile, Reconfigurable Manufacturing" was taught simultaneously here and at Wayne State University, and other

institutions expressed interest in the same course. We would like to add that in the April 2002 Edition of the Best Graduate School Rankings, pub­lished by US News and World Reports, the UM Mechanical Engineering graduate Program was ranked second, and "Industrial/ Manufacturing" Engineering was also ranked second among 145 US gradu­ate programs that provided data to calculate rankings. This is the highest ranking that the ME graduate pro­gram has received ever! There is no doubt that the reputation and accom­plishments of the ERC!R..\15 played a major role in this high ranking of the ME and the Manufacturing pro­grams. We are certain that this high ranking will attract even better stu­dents ro our graduate programs.

Worldwide Impact

Our presentation in China abour the ERC/RMS in 1997 influenced the cre­ation of a similar center there, and our 1999 ClRP keynote paper on "Reconfigurable Manufacturing Systems" served as a catalyst to start intensive research on reconfigurable manufacturing in Canada (London and Hamilton), Germany (Stuttgart), and Italy (Milan and Palermo). A keynote paper on R..\t1S and a plenary paper on R.JVI.S and RMTs were pre­sented at the 2000 Japan-USA Symposium and the 2000 Japan Machine Tool Show The CIRP " 1st International Conference on Reconfigurable Manufacturing," hosted by our Center in 2001, attract­ed 160 participants from 14 coun­tries, of which 80 were from industry (Canada, Germany, Japan, Italy, and the US). This conference has been a remarkable milestone that demon­strated the Center impact on acade­mia and industry. However, it is diffi­cult co evaluate the long-range impact of a new technology on the market­place and the educational infrastruc-

P R 0 C E S S E S

ture after only five and a half years of operation. We believe that RMS - a new research field initiated by us with obvious benefits to industry and soci­ety -will continue to grow over time. Tbis assertion is supported by a Delphi smdy, summarized in rhe U.S. National Research Cowtcil's Visionary Manufacturing Challenges for 2020 [4], which identified "adapt­able and reconfigurable systems'" as the number one priority technology in manufacturing for 2020. We are con­fident that only a few years from now the term "reconfigurable" will be used frequently both in manufactur­ing and other disciplines, and that sys­tem, machine, and control reconfig­urability wiU be basic requirements for new manufacturing systems. e

References 1. DeGaspari, J., 2002, "All in the family:

Flexible machining systems give manu­facturers a hedge on their bets." Mechanical Engineering Magazine, Vol. 124, no. 2.

2. Fabrication & Metalworking. Feb. 2002, pp. 27-29.

3. Koren, Y., Heisel. U., Jovane, F., Moriwaki, T., Pritchow, G., Ulsoy, A. G. and H. VanBrussel, "Reconfigurable Manufacturing Systems," CIRP Annals, Vol. 48, No.2, 1999, pp. 527-540 (keynote paper). CIRP keynote paper.

4. National Research Council, Visionary Manufacturing Challenges for 2020, National Academy Press, 1998.

5. Pande, P. and L. Holpp, "'What is Six Sigma," McGraw-Hill, NY, 2002. book.

6. Womack, J.P. and D.T. Jones, "Lean Thinking." Simon and Schuster, NY, 1996.

POWERTRAIN INTERNATIONAL 21